and Aplysina aerophoba
Carsten Thomsa, Matthias Wolffb, K. Padmakumarc, Rainer Ebela, and Peter Prokscha,*
a Institut für Pharmazeutische Biologie, Universität Düsseldorf, Universitätsstraße 1, Geb. 26.23, D-40225 Düsseldorf, Germany. Fax: +49-2 11-81-1 19 23.
E-mail: proksch@uni-duesseldorf.de
b Zentrum für Marine Tropenökologie, Fahrenheitstr. 6, D-28359 Bremen, Germany
c Department of Aquatic Biology and Fisheries, Kerala University Campus, Kariavattom, Trivandrum-695581, India
* Author for correspondence and reprint requests
Z. Naturforsch. 59 c, 113Ð122 (2004); received October 7/October 28, 2003
The Mediterranean sponges Aplysina aerophoba and A. cavernicola accumulate bromi- nated isoxazoline alkaloids including aplysinamisin-1 (1), aerophobin-2 (2), isofistularin-3 (3) or aerothionin (4) at concentrations up to 10% of their respective dry weights. In laboratory feeding experiments employing the polyphagous Mediterranean fish Blennius sphinx crude extracts of both Aplysina sponges were incorporated into artificial fish food at their physio- logical concentrations (based on volume) and offered to B. sphinx in choice feeding experi- ments against untreated control food. In addition to the Aplysina sponges, extracts from nine other frequently occurring Mediterranean sponges were likewise included into the experi- ments. Both Aplysina species elicited strong feeding deterrence compared to the other sponges tested. Bioassay-guided fractionation of A. cavernicola yielded the isoxazoline alka- loids aerothionin (4) and aplysinamisin-1 (1) as well as the 3,4-dihydroxyquinoline-2-carbo- xylic acid (8) as major deterrent constituents when tested at their physiological concentra- tions as present in sponges. Aeroplysinin-1 (5) and dienone (6), however, which are formed in A. aerophoba and A. cavernicola from isoxazoline precursors through bioconversion reac- tions upon tissue injury showed no or only little deterrent activity. Fractionation of a crude extract of A. aerophoba yielded aerophobin-2 (2) and isofistularin-3 (3) as major deterrent constituents against B. sphinx. We propose that the isoxazoline alkaloids 1Ð4 of Mediterra- nean Aplysina sponges as well as the 3,4-dihydroxyquinoline-2-carboxylic acid (8) (in the case of A. cavernicola) act as defensive metabolites against B. sphinx and possibly also against other predators while the antibiotically active bioconversion products aeroplysinin-1 (5) and dienone (6) may protect sponges from invasion of bacterial pathogens.
Key words: Chemical Defense, Fish Feeding Assay, Aplysina Sponges
Introduction
Marine sponges accumulate a large variety of structurally highly diverse secondary metabolites many of which are deterrent towards potential predators such as fishes (e.g. Pawlik et al., 1995;
Schupp et al., 1999; Becerro et al., 2003), possess anti-fouling activity (e.g. Thompson et al., 1985;
Martı´n and Uriz, 1993; Becerro et al., 1994) and/
or suppress the growth of other competing inver- tebrates (e.g. Sullivan et al., 1983; Porter and Targett, 1988; Turon et al., 1996). Among these various ecological roles, deterrence of potential predators appears to be most significant (e.g.
Braekman and Daloze, 1986; Proksch, 1999;
McClintock and Baker, 2001 and references cited therein). Taking into consideration that sponges
0939Ð5075/2004/0100Ð0113 $ 06.00 ”2004 Verlag der Zeitschrift für Naturforschung, Tübingen · http://www.znaturforsch.com ·D
often live in ecosystems such as coral reefs that are characterized by an exceptionally high feeding pressure (Carpenter, 1986) and that they further- more lack effective morphological defense mecha- nisms (Chanas and Pawlik, 1996), it is obvious that the frequent occurrence of deterrent and/or toxic metabolites in their tissues can be interpreted as chemical defense protecting these sessile and vul- nerable invertebrates from potential predators.
However, chemical defense of marine sponges through accumulation of bioactive natural pro- ducts is not only found in the tropics where feed- ing pressure by fishes is extreme but also in tem- perate marine habitats and even under the Antarctic ice cover (McClintock, 1987; Uriz et al., 1996; Becerro et al., 2003). Sponges of the order Verongida [which includes the genus Aplysina
(syn. Verongia)], for example, are rich in structur- ally unusual brominated isoxazoline alkaloids which are thought to be biogenetically derived from dibromotyrosine (Gopichand and Schmitz, 1979; Tymiak and Rinehart, 1981; Carney and Rinehart, 1995). Aplysina species occur in the Mediterranean Sea but also in the Atlantic Ocean (e.g. around the Canary Isles) and in the Carib- bean Sea (Pawlik et al., 1995) and usually contrib- ute to the dominating sponges present in the re- spective habitat. The alkaloid patterns of the two Mediterranean species A. aerophoba and A. caver- nicola share similarities including the presence of aerophobin-2 (2) and of aplysinamisin-1 (1) (which is not always present in A. aerophoba) (Teeyapant et al., 1993; Ciminello et al., 1997;
Thoms et al., 2003a, b). The assemblages of sec- ondary metabolites in both Aplysina sponges, however, differ mainly with regard to aerothionin (4) which occurs only in A. cavernicola and with regard to their yellowish pigments. The yellow color of A. aerophoba is due to the chemically highly labile pigment uranidine (7) which readily undergoes polymerization when exposed to air (Cimino et al., 1984) whereas A. cavernicola accu- mulates 3,4-dihydroxyquinoline-2-carboxylic acid (8) which is far more stable than uranidine (7).
Both species of Mediterranean Aplysina sponges feature an unusual wound-induced bio- conversion of brominated isoxazoline alkaloids which gives rise to the nitrile aeroplysinin-1 (5) which is in turn converted to dienone (6) (Teeya- pant and Proksch, 1993; Ebel et al., 1997). The bio- conversion products aeroplysinin-1 (5) and di- enone (6) were previously shown to be biologically strongly active against marine bacteria and algae whereas their isoxazoline precursors were not (Teeyapant et al., 1993; Weiss et al., 1996). So far, however, no detailed studies on the fish deterrent properties of isolated natural products present in A. aerophoba and A. cavernicola have been con- ducted. Since both Aplysina species are among the most common sponges found in the Mediterra- nean Sea (especially A. aerophoba) and are appa- rently not preyed upon by fishes in their natural habitat we initiated a detailed study on the fish deterrent properties of the crude extracts and of the major secondary metabolites of A. aerophoba and A. cavernicola using the polyphagous Medi- terranean fish Blennius sphinx as a model organ- ism.
Material and Methods
Collection of sponge material and preparation of crude extracts
Specimens of ten different sponge species in- cluding A. cavernicola were collected by SCUBA diving off the coast of Elba in the Mediterranean Sea in April 2000. All samples were kept sub- merged in zip-lock bags filled with seawater during transportation to the laboratory. Before preserv- ing them in EtOH, weight and volume (measured by displacement of water) of all collected sponges were determined. Following determination of their volumes all sponge samples were minced with a blender and subsequently extracted at room tem- perature exhaustively in MeOH and CH2Cl2. EtOH, MeOH and CH2Cl2 extracts were com- bined to give crude extracts which in the following are referred to as extracts from EtOH preserved sponges. The collected biomass of A. cavernicola was divided into two parts. One part was extracted as explained above whereas the second part of the tissue was lyophilized, ground with a mortar and extracted exhaustively with MeOH and CH2Cl2. Specimens of the sponge Aplysina aerophoba that had been collected at Rovinj, Croatia were like- wise lyophilized and extracted. The MeOH and CH2Cl2extracts were subsequently combined to give crude extracts which in the following are re- ferred to as extracts from lyophilized sponges.
Fractionation of the crude extracts from A. aerophoba and A. cavernicola
The crude extracts of lyophilized A. aerophoba or A. cavernicola were taken to dryness and par- titioned between EtOAc and H2O. 3,4-Dihy- droxyquinoline-2-carboxylic acid (8) was mainly detected in the aqueous phase resulting from par- titioning of the A. cavernicola extract. It was iso- lated and purified by repeated gel permeation chromatography over LH-20 Sephadex (Amer- sham Pharmacia Biotech, Freiburg, Germany) using MeOH as a solvent. Compound identifica- tion was based on comparison of NMR spectra with previously described data (Fattorusso et al., 1971; this study describes 3,4-dihydroxyquinoline- 2-carboxylic acid as derived from the sponge A. aerophoba instead of from A. cavernicola; later the collected sponge proved to be A. cavernicola).
The EtOAc phases resulting from partitioning of the crude extracts of lyophilized A. aerophoba and A. cavernicola were also taken to dryness and
further partitioned between MeOH/H2O and cy- clohexane followed by a separation of the MeOH/
H2O layers by column chromatography on Sepha- dex LH-20 using MeOH as a solvent. Further puri- fication was achieved by repeated column chroma- tography on silica gel using various mixtures of CH2Cl2 and 2-propanol as solvent systems. Frac- tions were monitored by TLC on pre-coated silica gel plates (Merck, Darmstadt, Germany) employ- ing the same solvent systems and by HPLC analy- sis using an HPLC system coupled to a photo- diode-array detector (Dionex, Germany). Routine detection was at 254 nm. The separation column (125 ¥ 4 mm i.d.) was prefilled with Eurosphere C-18 (5µm) (Knauer, Germany). A solvent system consisting of 0.02% phosphoric acid at pH 2 and MeOH at a gradient increasing linearly from 10%
MeOH to 100% within 35 min was used for com- pound separation. Identification of compounds isolated from both Aplysina sponges was based on their online UV spectra, on HPLC/ESIMS data and on direct comparison with previously isolated standards (Ebel et al., 1997). All compounds were quantified by HPLC using calibration curves ob- tained for the respective isolated natural products.
For isolation of aeroplysinin-1 (5) and dienone (6) fresh sponge biomass (containing H2O) that had been immersed into EtOH immediately after collection was used. Compounds were isolated, identified and quantified as described above for the isoxazoline alkaloids isolated from lyophilized A. cavernicola material.
Fish feeding bioassays
For feeding experiments with crude sponge ex- tracts and/or fractions derived thereof the respec- tive amounts (in mg) obtained following extrac- tion of 10 ml of EtOH preserved sponge tissue were incorporated into the same volume of artifi- cial fish food. For assays involving isolated pure metabolites of A. cavernicola and A. aerophoba quantification was performed by HPLC as de- scribed above. Aerothionin (4), aplysinamisin-1 (1) and isofistularin-3 (3) (precursors of the injury- induced bioconversion) as well as compound 8 were quantified in lyophilized sponge tissue. For quantification of the bioconversion products aero- plysinin-1 (5) and dienone (6) EtOH preserved material of fresh A. cavernicola was used. Artifi- cial fish food was treated with physiological con- centrations of secondary sponge metabolites as de-
termined in 10 ml EtOH preserved sponge material. In addition to experiments with crude extracts, fractions and pure metabolites, feeding assays with mixtures of pure sponge compounds [(3,4-dihydroxyquinoline-2-carboxylic acid + aero- thionin) and (3,4-dihydroxyquinoline-2-carboxylic acid + dienone)] were also performed. In these lat- ter bioassays each compound tested was incorpo- rated into fish food at its respective physiological concentration as quantified in 10 ml of sponge tis- sue.
The fish feeding assays performed in this study were modeled based on a method previously de- scribed by Hay et al. (1998). This method had orig- inally been designed for feeding experiments with sea urchins but was later successfully adapted by Schupp et al. (1999) for fish feeding assays. All metabolites used for feeding experiments were dissolved in MeOH and incorporated into 1.053 g of ground commercial fish food granules (Tetra- Marin Granulat, TetraWerke, Melle, Germany).
This mixture was then added to 9 ml of 2% molten agar at 48∞C to give an overall volume of fish food of 10 ml. Thus the concentrations of extracts, frac- tions and pure compounds used in the bioassays matched the natural concentrations as found in the sponges. Aeroplysinin-1 (5) was incorporated into the artificial fish food at the molar concentration as quantified for dienone (6) in the EtOH pre- served A. cavernicola material. Control food was likewise prepared but food granules were mixed only with MeOH without sponge metabolites. All freshly prepared food mixtures (with or without sponge metabolites) were then poured into two rectangular molds (2 mm ¥ 25 mm ¥ 250 mm) placed on a window screen mesh. One of the molds was filled with treated, the other one with control food. When the agar had cooled com- pletely, it became firmly attached to the window screen. By cutting it perpendicular to the food pieces, six equivalent food strips were obtained, each of them containing one rectangle with treated food and one control food piece (each covering an area of 10¥15 window screen squares) (Fig. 1).
Feeding experiments were performed in an aquarium (size: 170 cm¥80 cm¥40 cm) using 70 individuals of the polyphagous Mediterranean fish Blennius sphinx. Food strips containing one treated and one control food rectangle were placed at random positions on the bottom of the aquarium. Six parallel experiments (with one sponge extract or fraction, respectively) or less
Fig. 1. Food strips containing agar based control food pieces and food pieces treated with sponge metabolites.
The picture was taken after preliminary feeding assays employing different sponge extracts.
were usually carried out per day. The food strips were checked at regular time intervals and were removed when approximately a third part of the total food mass had been consumed. As each fish bite made up only for a minute amount of agar based food, few trial bites at strongly feeding de-
0
20 40 60 80 1 0 0 1 2 0
A. cavernicola lyoph.A. cavernicola EtOHA. aerophoba lyoph.
Agelas oroides Axinella damicornisHam
igera hamigera
Acanthella acuta Ircinia fasciculata Spongia officinalis Chondrosia reniformis
Ircinia spinosula Petrosia ficiformis n = 17
p < 0.001 n = 17 p < 0.001
Fractionof total amountoffoodeaten (%)
control treated
0 20 40 60 80 100
n = 6 p < 0.001
n = 12 p < 0.001
n = 6 p < 0.001
n = 12 p < 0.001
n = 12 p < 0.001
n = 18 p < 0.001
n = 12 p < 0.001
n = 12 p = 0.051
n = 12 p = 0.053
n = 12 p = 0.081
Fig. 2. Feeding experiments with crude extracts from 11 Mediterranean sponge species towards Blennius sphinx.
Crude extract from A. cavernicola was derived from lyophilized (lyoph.) and from EtOH preserved (EtOH) fresh sponge tissue. Crude extract from A. aerophoba was derived only from lyophilized, all other sponges only from EtOH preserved material.
terrent food were almost not visible. The removal of one third of food treated with an edible extract or of control food corresponded to approximately 100 fish bites. Individual experiments lasted be- tween one and two hours. To determine the amount of control and treated food eaten, the number of empty squares in the window screen was counted (Hay et al., 1998). The results ob- tained from several replicate experiments were pooled and analyzed with a paired t-test.
Determination of nitrogen and carbon content in sponge tissues and in the artificial fish food granules
Three pieces of each sponge species were lyo- philized and ground with a mortar. About 0.3 g of each sample were then measured with a Induc- tively Coupled Plasma Emission Spectroscope (Jobin Yvon/HORIBA Ltd.). Carbon and nitrogen content in the sponge tissues were determined as percentages of their dry weights in order to facili- tate a comparison with the nutrient values of the commercial fish food granules used for the feed- ing experiments.
Results
Intensity of feeding deterrence of crude extracts from eleven Mediterranean sponge species
The results of the feeding experiments with crude extracts of 11 Mediterranean sponges are shown in Fig. 2. Nine of the analyzed extracts proved to be significantly repellent against the fishes. While food treated with crude extracts from Aplysina cavernicola, A. aerophoba, Agelas oroi- des, and Axinella damicornis remained almost untouched by B. sphinx, food containing extracts of Hamigera hamigera, Acanthella acuta, Ircinia fasciculata, and Spongia officinalis was consumed more often but still significantly less than the un- treated control food (Fig. 2). There was no clear difference in feeding deterrence between the ex- tract from lyophilized A. cavernicola and the one
1
NH
O O
N
N H
N NH2 O
H Br
H3CO Br
O H Br
H3CO Br
Br
Br
OCH3 N OH
H O
O N N
H
O O
N
4
5
8 N
H O
OH
OH NH O
O
N O
O N Br
Br O
H Br
H3CO Br
Br
Br
OCH3 OH
N OH
OH
COOH 3
2
OCH3
Br Br
O H
OH NC
O
Br Br
OH H2NOC
6
N H
OH O
OH
OH
7 N
H
O O
N
NH
N NH2 O
H Br
H3CO Br
Fig. 3. Structures of Aplysina alkaloids: aplysinamisin-1 (1), aerophobin-2 (2), isofistularin-3 (3), aerothionin (4), aeroplysinin-1 (5), dienone (6), uranidine (7), and 3,4- dihydroxyquinoline-2-carboxylic acid (8).
obtained by extraction of fresh sponge tissue. Arti- ficial fish food treated with extracts from Chon- drosia reniformis, Ircinia spinosula, and Petrosia ficiformis on the other hand had no significant de- terrent effects on the fishes.
Intensity of feeding deterrence of pure metabolites from A. cavernicola and A. aerophoba
Bioassay-guided fractionation of the crude ex- tract of lyophilized tissue of A. cavernicola using the above described feeding bioassay with B.
sphinx resulted in the isolation of aplysinamisin-1 (1), aerophobin-2 (2), aerothionin (4) and of 3,4- dihydroxyquinoline-2-carboxylic acid (8) as deter- rent secondary metabolites (Fig. 3). All of these sponge metabolites when tested at their natural concentrations proved to be highly deterrent to B. sphinx with very few fish bites observed at the artificial food pellets (Fig. 4). When the major bro- minated isoxazoline alkaloid aerothionin (4) and 3,4-dihydroxyquinoline-2-carboxylic acid (8) were jointly incorporated into the artificial food pellets at their respective natural concentrations an in- crease of feeding deterrence of this natural mix- ture towards B. sphinx over the individual respec- tive metabolites was observed (Fig. 4).
An extract prepared from EtOH preserved fresh A. cavernicola proved to be almost deterrent towards B. sphinx as the one prepared from lyo- philized sponge tissue (Fig. 2) even though HPLC analysis revealed both extracts to be very different with regard to their major secondary metabolites.
Whereas the extract from lyophilized A. caverni- cola featured aerothionin (4), aplysinamisin-1 (1), aerophobin-2 (2) and 8 as major UV-absorbing compounds the extract prepared from EtOH pre- served fresh sponge tissue was characterized by two major peaks which included 8 and dienone (6). Origin of the latter through bioconversion of isoxazoline precursors in aqueous media contain- ing organic solvents has been demonstrated before (Ebel et al., 1997). A known intermediate of this bioconversion of isoxazoline alkaloids is the nitrile aeroplysinin-1 (5) (Fig. 3) (Ebel et al., 1997). When the bioconversion products aeroplysinin-1 (5) and dienone (6) were incorporated into artificial fish food (each tested at the concentration of the latter compound as elucidated through HPLC analysis of the extract prepared from EtOH preserved A. cavernicola) aeroplysinin-1 (5) proved to be weakly (compared for example to 3 or 4) but nev-
0,0 20,0 40,0 60,0 80,0 100,0 120,0
1 2 3 4 5 6 7 8 9
control treated
Fractionof total amountoffoodeaten (%)
0 20 40 60 80 100
Aplysinamisin-1 (1) Aerophobin-2 (2) Isofistularin-3 (3) Aerothionin (4) Compound 8 Aeroplysinin-1 (5) Dienone (6) Compound8 + Aerothionin(4) Compound8 + Dienone (6)
n = 6 p < 0.001
n = 12 p < 0.001 n = 6
p < 0.001 n = 6 p < 0.001
n = 12 p < 0.001
n = 11 p < 0.001
n = 12 p = 0.065
n = 6 p < 0.001
n = 6 p < 0.001
I II III
Fig. 4. Feeding experiments with isolated metabolites from Aplysina sponges towards Blennius sphinx.
(I) Metabolites isolated from lyophilized sponge tissue: Aerothionin (4), aplysinamisin-1 (1), and 3,4-dihydroxy- quinoline-2-carboxylic acid (8) were tested in concentrations as determined in A. cavernicola, isofistularin-3 (3), and aerophobin-2 (2) as determined in A. aerophoba tissue.
(II) Metabolites isolated from EtOH preserved tissue: Dienone (6) was tested in concentrations as determined in A. cavernicola. Aeroplysinin-1 (5) was tested at the same molar concentration as dienone (6).
(III) Mixtures of 3,4-dihydroxyquinoline-2-carboxylic acid (8) with other pure metabolites isolated from A. caverni- cola tissue.
ertheless significantly deterrent towards B. sphinx whereas dienone (6) had no feeding deterrent ef- fect at the chosen concentration (Fig. 4). When dienone (6) and compound 8 isolated from A. ca- vernicola were jointly added to artificial fish food and presented to B. sphinx the intensity of feeding deterrence was comparable to that of 8 alone (Fig. 4). Thus there was no significant additive ef- fect caused by the presence of dienone (6), while in feeding experiments with compound 8 plus aer- othionin (4) an enhanced deterrent effect was ob- served (Fig. 4).
The extract of lyophilized tissue of A. aero- phoba which had likewise proven to be highly deterrent towards B. sphinx contained isofistu- larin-3 (3) and aerophobin-2 (2) as major bromi- nated isoxazoline alkaloids as indicated by HPLC analysis (Fig. 3). In addition traces of the chemi- cally unstable pigment uranidine (7) which readily polymerizes when exposed to oxygen were de- tected. Due to the instability of uranidine which prevented accurate quantification in the sponge
tissue this compound was not included into the fish feeding bioassays. Isofistularin-3 (3) and aero- phobin-2 (2), however, just like the brominated isoxazoline alkaloids isolated from A. cavernicola (1 and 4) proved to be highly deterrent towards B.
sphinx when added to artificial fish food at its nat- ural concentration as present in lyophilized sponge tissue (Fig. 4).
Nitrogen and carbon content in analyzed sponges and artificial fish food
For an assessment of the nutritional values of the sponge tissues and for comparison with the ar- tificial fish food used in this study nitrogen and carbon contents were measured. The data re- vealed that the nitrogen fractions in the sponge samples vary from 5.05% of dry weight as was the case for H. hamigera to 11.02% as found for A. oroides. Carbon contents of the analyzed sponges ranged from 25.30% for A. damicornis to 39.42% as found for I. spinosula. Nitrogen and
carbon contents of the artificial fish food em- ployed in this study (7.47 and 45.03%, respec- tively) were almost in the same ranges as found in the sponge specimens.
Discussion
The remarkable chemical diversity and biologi- cal activity of natural products found in marine sponges is often explained in terms of chemical defense of these sessile soft-bodied invertebrates against various biotic stress factors such as preda- tion, allelopathy and biofouling (e.g. Braekman and Daloze, 1986; Uriz et al., 1991; Proksch, 1999;
McClintock and Baker, 2001). Among the dif- ferent stress factors that affect and influence their fitness and ecological success, predation by fishes is probably the most important.
The nutrient analysis of eleven different sponges from the Mediterranean Sea performed in this study revealed similar carbon and nitrogen contents as in the artificial fish food used for the feeding experiments indicating that these sponges could be a valuable food source to fishes. How- ever, the fact that sizeable tissue damage of sponges due to feeding by fishes is usually rare in their natural environment (Green, 1977) points towards an effective defense of sponges against predatory fishes.
In this study we have focused on the chemical defense of Mediterranean Aplysina sponges against fishes. We selected Blennius sphinx as a test organism for our feeding bioassays as this small fish (about 5 cm in length) is (a) very abun- dant at the coastlines of the Mediterranean Sea, (b) polyphagous, feeding on algae as well as on invertebrates in its natural environment by grazing on rock surfaces and (c) inhabits the same ecosys- tem as the sponge species selected for our study (Riedl, 1983). Thus, we consider B. sphinx as a po- tential predator of Mediterranean sponges. By using a feeding assay similar to that originally de- scribed by Hay et al. (1998) (placing the experi- mental fish food at the bottom of the aquarium) we intended to simulate the natural feeding condi- tions of the test fishes. In order to minimize loss of compounds due to leaching into the surround- ing sea water during the feeding experiments we minimized exposure time of the food pieces in the aquarium by using a large number of test fishes to obtain test results within a rather short time of 1Ð2 h or even less in case of inactive components.
As part of our feeding studies we offered ex- tracts of eleven common sponge species from the Mediterranean Sea to fishes of the species B.
sphinx which were kept in an aquarium. The re- sults obtained in this experiment (Fig. 2) were re- markably similar to findings recently reported by Becerro et al. (2003) who tested the deterrent properties of sponge extracts against fishes under field conditions. In agreement with our results Becerro et al. (2003) found Aplysina aerophoba, Agelas oroides and Axinella damicornis to rank amongst those sponges that are well protected against small Mediterranean fishes living close to the substrate. As found in our study, extracts from Ircinia fasciculata and Petrosia ficiformis were likewise significantly avoided when compared to controls. However, fishes fed on them more often during the experiments as on artificial food con- taining extracts of the former three sponges. This similarity of the results obtained in our feeding experiments with B. sphinx compared to the re- sults obtained by Becerro et al. (2003) under field conditions indicates that the data obtained by the laboratory test system with B. sphinx has ecologi- cal relevance.
Our feeding experiments with B. sphinx indi- cated that both Aplysina sponges contain highly deterrent secondary metabolites (Fig. 2). This finding is in agreement with the results of a study by Pawlik et al. (1995) who analyzed 71 Caribbean sponges for fish deterrence and reported five Aplysina species (not including the species studied in this report) to rank among the chemically best protected sponges encountered.
In our assays the extract obtained from lyophi- lized A. cavernicola was comparable with regard to deterrency to the extract that had been pre- pared by immersing fresh sponge into EtOH (Fig. 2) even though both extracts clearly differ with regard to their major secondary metabolites.
Whereas the former is characterized by the pres- ence of several brominated isoxazoline alkaloids such as aplysinamisin-1 (1), aerophobin-2 (2) and aerothionin (4) as well as copious amounts of 8 the latter features compound 8 and the dienone (6) as major secondary metabolites with isoxazo- line alkaloids present only in trace amounts. Ebel et al. (1997) had shown previously that isoxazoline alkaloids can be converted in vitro into the nitrile aeroplysinin-1 (5) which in turn gives rise to die- none (6) using a cell free extract obtained from A. aerophoba or A. cavernicola. Furthermore it
had been suggested that similar reactions occur also in situ upon wounding of sponge tissue (Ebel et al., 1997) or when fresh sponges are extracted in presence of water (Teeyapant and Proksch, 1993).
Compound 8, which is clearly no isoxazoline al- kaloid, remains apparently unaffected by these conversions and is detected in extracts of A. caver- nicola regardless of the method chosen for extrac- tion. In this context, it should be noted that Puyana et al. (2003) found no evidence for a sim- ilar conversion of isoxazoline alkaloids in the Car- ibbean sponges Aplysina insularis and A. archeri.
Through bioassay guided fractionation of an ex- tract obtained from lyophilized A. cavernicola aerothionin (4) and aplysinamisin-1 (1) as well as 3,4-dihydroxyquinoline-2-carboxylic acid (8) were found to be the major deterrent secondary metab- olites. Each compound when tested at its physio- logical concentration as present in the sponges sig- nificantly deterred feeding by B. sphinx (Fig. 4).
When pieces of artificial fish food were treated with a mixture of compound 8 and aerothionin (4) an additive or synergistic effect with regard to feeding deterrence was found (Fig. 4).
When the extract prepared from fresh A. caver- nicola that had been immersed into EtOH was subjected to a bioassay-guided fractionation only compound 8 was identified as defensive metabo- lite (Fig. 4). The dienone (6) on the other hand had no significant fish deterrent activity (Fig. 4).
Adding dienone (6) together with compound 8 to artificial food at their respective physiological concentrations the feeding deterrent effect of the treated fish food increased only to a small extent when compared to fish food that had been treated only with compound 8 (Fig. 4). In contrast to the dienone (6), which apparently had no significant feeding deterrent activity, aeroplysinin-1 (5) when tested at the concentration of the dienone (6) caused a weak (compared for example to 3 or 4) but nevertheless statistically significant feeding in- hibition of B. sphinx (Fig. 4).
Isofistularin-3 (3) and aerophobin-2 (2) are among the major isoxazoline alkaloids present in the sibling sponge species A. aerophoba. When the latter alkaloids were incorporated into artificial fish food at the concentrations as found in A. aero-
phoba and offered to B. sphinx both isofistularin- 3 (3) and aerophobin-2 (2) provoked pronounced antifeedant activity (Fig. 4). Direct comparison of the antifeedant activity of the various isoxazoline alkaloids 1Ð4 from A. aerophoba and A. caverni- cola towards B. sphinx with regard to possible structure-activity relationships, however, is not possible, since the respective compounds were not tested at equimolar but rather at their physiologi- cal concentrations as detected in both Aplysina sponges.
Based on the results obtained in our study we propose that the brominated isoxazoline alkaloids of A. aerophoba and A. cavernicola as well as com- pound 8 that is present only in A. cavernicola act as constitutive defense compounds against poten- tial fish predators. Aeroplysinin-1 (5) and dienone (6), on the other hand, are clearly inferior in com- parison with regard to inhibiting feeding of B.
sphinx. Nevertheless they are characterized by pronounced antibiotic properties against a multi- tude of terrestrial as well as marine bacteria (Teeyapant and Proksch, 1993; Teeyapant et al., 1993; Weiss et al., 1996; Debitus et al., 1998). Inter- estingly, isoxazoline alkaloids analyzed in these studies (e.g. 1Ð4) are devoid of antibacterial activ- ity when tested against the same range of bacteria (Teeyapant et al., 1993; Weiss et al., 1996). Thus, one may speculate that aeroplysinin-1 (5) and di- enone (6) protect the sponges primarily from bac- terial invasion (e.g. following wounding) whereas their isoxazoline precursors serve as chemical de- fense metabolites against fishes such as B. sphinx.
Acknowledgements
We gratefully acknowledge Prof. W. Kaiser for the analysis of nutrient contents in the sponge samples, our colleagues Gero Eck and Jan Hiort for diving assistance, Wafaa Hassan for her help while carrying out feeding experiments, Sabine Borstel for technical assistance in isolating the compounds and the staff of the Hydra-Institut für Meereswissenschaften at Elba, Italy for diving and technical support. This work was generously sup- ported by grants to P. P. (BMBF and by the Fonds der Chemischen Industrie).
Becerro M. A., Turon X., Uriz M. J., and Lopez N. I. McClintock J. B., Swenson D., Trapido-Rosenthal H., (1994), Antimicrobial activity and surface bacterial and Banghart L. (1997), Ichthyodeterrent properties film in marine sponges. J. Exp. Mar. Biol. Ecol. 179, of lipophilic extracts from Bermudian sponges. J.
195Ð205. Chem. Ecol. 23, 1607Ð1620.
Becerro M. A., Thacker R. W., Turon X., Uriz M. J., and Pawlik J. R., Chanas B., Toonen R. J., and Fenical W.
Paul V. J. (2003), Biogeography of sponge chemical (1995), Defenses of Caribbean sponges against preda- ecology: comparisons of tropical and temperate de- tory reef fish. I. Chemical deterrency. Mar. Ecol. Prog.
fenses. Oecologia 135, 91Ð101. Ser. 127, 183Ð194.
Braekman J. C. and Daloze D. (1986), Chemical defense Porter J. W. and Targett N. M. (1988), Allelochemical in- in sponges. Pure Appl. Chem. 58, 357Ð364. teractions between sponges and corals. Biol. Bull. 175, Carney J. R., and Rinehart K. L. (1995), Biosynthesis of 230Ð239.
brominated tyrosine metabolites by Aplysina fistu- Proksch P. (1999), Chemical defense in marine ecosys- laris. J. Nat. Prod. 58, 971Ð985. tems. In: Functions of Plant Secondary Metabolites Carpenter R. C. (1986), Partitioning herbivory and its and their Exploitation in Biotechnology (Wink M.,
effects on coral algal communities. Ecol. Monogr. 56, ed.). Academic, Sheffield, pp. 134Ð154.
345Ð365. Puyana M., Fenical W., and Pawlik J. R. (2003), Are
Chanas B. and Pawlik J. R. (1996), Does the skeleton of there activated chemical defenses in sponges of the a sponge provide a defense against predatory reef genus Aplysina from the Caribbean? Mar. Ecol. Prog.
fish? Oecologia 107, 225Ð231. Ser. 246, 127Ð135.
Ciminello P., Fattorusso E., Foriono M., Magno S., and Riedl H. (1983), Fauna und Flora des Mittelmeeres. Ver- Pansini M. (1997), Chemistry of Verongida Sponges lag Paul Parey, Hamburg.
VIII Ð Bromocompounds from the Mediterranean Schupp P., Eder C., Paul V., and Proksch P. (1999), Dis- Sponges Aplysina aerophoba and Aplysina caverni- tribution of secondary metabolites in the sponge cola. Tetrahedron 53, 6565Ð6572. Oceanapia sp., and its ecological implications. Mar.
Cimino G., de Rosa S., de Stefano S., Spinella A., and Biol. 135, 573Ð580.
Sodano G. (1984), The zoochrome of the sponge Ve- Sullivan B., Faulkner D. J., and Webb L. (1983), Sipho- rongia aerophoba (“Uranidine”). Tetrahedron Lett. nodictidine, a metabolite of the burrowing sponge
25, 2925Ð2928. Siphonodictyon sp. that inhibits coral growth. Science
Debitus C., Guella G., Mancini I., Waikedre J., Guemas 221, 1175Ð1176.
J.-P., Nicolas J.-L., and Pietra F. (1998), Quinolones Teeyapant R. and Proksch P. (1993), Biotransformation from a bacterium and tyrosine metabolites from its of brominated compounds in the marine sponge Ve- host sponge, Suberea creba from the Coral Sea. J. Mar. rongia aerophoba. Evidence for an induced chemical Biotechnol. 6, 136Ð141. defense? Naturwissenschaften 80, 369Ð370.
Ebel R., Brenzinger M., Kunze A., Gross H. J., and Teeyapant R., Kreis P., Wray V., Witte L., and Proksch Proksch P. (1997), Wound activation of protoxins in P. (1993), Brominated secondary compounds from the marine sponge Aplysina aerophoba. J. Chem. Ecol. 23, marine sponge Verongia aerophoba and the sponge
1451Ð1462. feeding gastropod Tylodina perversa. Z. Naturforsch.
Fattorusso E., Forenza S., Minale L., and Sodano G. 48c, 640Ð644.
(1971), Isolation of 3,4-dihydroxyquinoline-2-carbox- Thompson J. E., Walker N. D., and Faulkner D. J. (1985), ylic acid from the sponge Aplysina aerophoba. Gazz. Screening and bioassays for biologically active sub- Chim. Ital. 101, 104Ð105. stances from forty marine sponge species from San Gopichand Y. and Schmitz F. J. (1979), Marine natural Diego, California, USA. Mar. Biol. 88, 11Ð21.
products: fistularin-1, -2 and -3 from the sponge Thoms C., Ebel R., Hentschel U., and Proksch P.
Aplysina fistularis forma fulva. Tetrahedron Lett. 41, (2003a), Sequestration of dietary alkaloids by the
3921Ð3924. spongivorous marine mollusc Tylodina perversa. Z.
Green G. (1977), Ecology of toxicity in marine sponges. Naturforsch. 58c, 426Ð432.
Mar. Biol. 40, 207Ð215. Thoms C., Horn M., Wagner M., Hentschel U., and Hay M. E., Stachowicz J. J., Cruz-Rivera E., Bullard S., Proksch P. (2003b), Monitoring microbial diversity Deal M. S., and Linquist N. (1998), Bioassays with ma- and natural products profiles of the sponge Aplysina rine and freshwater macroorganisms. In: Methods in cavernicola following transplantation. Mar. Biol. 142, Chemical Ecology, Vol. 2, Bioassay Methods (Haynes 685Ð692.
K. F. and Millar J. G., eds.). Chapman and Hall, New Turon X., Becerro M. A., and Uriz M. J. (1996), Seasonal
York, pp. 39Ð141. patterns of toxicity in benthic invertebrates: The
Martı´n D. and Uriz M. J. (1993), Chemical bioactivity of encrusting sponge Crambe crambe (Poecilosclerida).
Mediterranean benthic organisms against embryos Oikos 75, 33Ð40.
and larvae of marine invertebrates. J. Exp. Mar. Biol. Tymiak A. A. and Rinehart Jr. K. L. (1981), Biosynthesis
Ecol. 173, 11Ð27. of dibromotyrosine-derived antimicrobial compounds
McClintock J. B. (1987), Investigation of the relationship by the marine sponge Aplysina fistularis. J. Am.
between invertebrate predation and biochemical com- Chem. Soc. 103, 6763Ð6765.
position, energy content, spicule armament and toxic- Uriz M. J., Martin D., Turon X., Ballesteros E., Hughes ity of benthic sponges at McMurdo Sound, Antarctica. R., and Acebal C. (1991), An approach to the ecologi- Mar. Biol. 94, 479Ð487. cal significance of chemically mediated bioactivity in McClintock J. B. and Baker B. J. (2001), Marine Chemi- Mediterranean benthic communities. Mar. Ecol. Prog.
cal Ecology. CRC, Boca Raton, Florida. Ser. 70, 175Ð188.
Uriz M. J., Becerro M. A., Tur J. M., and Turon X. sponge Aplysina fistularis. II. Chemical evidence. Mar.
(1996), Location of toxicity within the Mediterranean Biol. 88, 27Ð32.
sponge Crambe crambe (Demospongiae: Poeciloscler- Weiss B., Ebel R., Elbrächter M., Kirchner M., and ida). Mar. Biol. 124, 583Ð590. Proksch P. (1996), Defense metabolites from the ma- Walker R. P., Thompson J. E., and Faulkner D. J. (1985), rine sponge Verongia aerophoba. Biochem. Syst. Ecol.
Exudation of biologically active metabolites in the 24, 1Ð12.
